A Fingerprint Pattern Approach to Hill Cipher

Implementation
Andysah Putera Utama Siahaan
Faculty of Computer Science
Universitas Pembangunan Panca Budi
Jl. Jend. Gatot Subroto Km. 4,5 Sei Sikambing, 20122, Medan, Sumatera Utara, Indonesia
andiesiahaan@gmail.com
Abstract - The human fingerprint always becomes the way
to verify the originality of the ownership. It can be
connected to the security methods to increase the security
level. Hill Cipher in one of the cryptography algorithms
that can attach the digital fingerprint pattern. There are
several matrix sizes to implement its process. This study
focuses to a 3x3 matrix in the application. We need to
provide nine integer numbers to perform the encryption
which determinant has already been tested before. We try
to link the digital fingerprint pattern to produce the
automatic key generator. Not all the determinant value can
get the ciphertext back to the original message. A threshold
is used to adjust the determinant. It produces the different
numbers when to be shifted. The correct numbers will be
occupied in the matrix. When the numbers are available,
the cryptography process can be performed.

II. THEORIES
A. Digital Fingerprint
A pattern of the fingerprint is generated when the finger is
pressed against a smooth surface. The most obvious structural
characteristics of the fingerprint are ridges and valleys. They
often run parallel, but they may bifurcate or terminate
unexpectedly sometimes [3]. Fingerprints are the curved
pattern in the skin fingertips. These streaks are not similar
among human being although they are identical. So that
fingerprints can be used to identify a person validity. Electronic
scanners capture the fingerprints based on light reflections of
the finger's ridges and valleys, ultrasonics, or the electrical
properties of the finger's ridges and valleys [4].

Keywords - Fingerprint, Hill Cipher, Cryptography, Security

I. INTRODUCTION
Fingerprint technique is one of the most important
authentication technologies based on the pattern of ridges and
valleys. Various applications are using this technology, but
the security problems are still not fully solved [3]. Hill cipher
uses a matrix to calulate the plaintext and produce the
ciphertext. It contains nine cells. Each cell is given the integer
number. These can be generated randomly. We try to propose
the producing technique. The digital fingerprint pattern is split
into nine segments. After the image is getting thresholded, it
forms a binary pattern. Each segment consists of the value of
"0" and "1" where "0" represents the white color and "1" to the
black or vise verse. It depends on the commitment. The total of
the value becomes the number inserted to each cell in matrix.
The threshold becomes the secondary key to produces the same
numbers when decrypting the ciphertext.
In the digital era, the picture can be used as an evidence or a
key [2]. We propose to combine the static image to authenticate
the user in Hill Cipher. Static, here, means the fingerprint is not
captured in real time, but it has already captured as a picture.
We do not collect the fingerprint samples. The image is scaled
to a specific dimension. Several image processing steps are
needed to process the bitmap into the binary data. This project
alternatively aims to generate numbers without calculate the
determinant manually.

Figure 1. Ridges and Valleys

Figure 1 show the ridges and valleys on a fingerprint. The
fingerprint of every people is different. The fingerprint pattern
has a different region among the model. It creates special
shapes respectively. These areas may be classified into three
models such as loop, delta, and whorl (Figure 2).

Figure 2. Special regions and core points

B. Hill Cipher
Hill Cipher is one of the symmetric key cryptographic
algorithms in data encryption [1]. To avoid the key matrix is
not invertible, matrix keys are generated using a binomial
coefficients newton. The encryption process and descriptions
use the same key. The plaintext itself may use image, audio or
text. Hill Cipher algorithm uses a matrix of m x m as a key for
encryption and decryption. The basic theory of matrix used in
Hill Cipher include the multiplication between the matrix and
perform the inverse of the matrix. It is an application of
cryptography modulo arithmetic.
Hill Cipher invented by Lester S. Hill in 1929 [5]. This
cryptographic technique was created to produce the cipher text
that cannot be solved by using frequency analysis techniques.
It does not replace the plaintext alphabet with another for the
same ciphertext. Hill Cipher is a block cipher as well. The
plaintext will be divided into several blocks of a certain size.
Each character in the block will affect the other characters in
the encryption and decryption process. The same characters in
plaintext are not mapped to the same character in the ciphertext.
Hill Cipher is the classic cryptographic algorithm. It is very
difficult to be solved by cryptanalyst if we have the ciphertext
only. However, this technique can be solved quite easily when
the cryptanalyst has the ciphertext file and a plaintext file
chunks. Cryptanalysis technique is called known-plaintext
attack [6].

The image needs to be converted first. After the image has

been in the grayscale mode, we need to convert it back to the
black and white. In this process, there a value to limit the color
intensity. It is called the threshold. After the processes are done,
the segmentation is carried out. Figure 4 illustrates the
segmentation after the image preprocessing. The amount of the
pixel in every segment is entirely different because the shape of
the finger is oval.

Figure 4. The fingerprint segmentation

Now we have nine segment. There are only two colors in this
stage such as red and white since we have limited the color
intensity before.

III. PROPOSED WORK

In this step, we try to figure out how the method works. We
use the real fingerprint scanned. The fingerprint image must be
converted into a grayscale picture because we do not use athree-layer color intensity. It must be cleaned from noise. To do
this, we have to convert the grayscale into a black and white.
Afterward, we divide the image into nine segments. The limit
of the segment is up to the commitment. Since the fingerprint
image is not square, the segmentation is not square as well. The
dimension of the image is 120 x 180 pixels. The following
figure is the image taken with a scanner. It is still original. If
we see the color, it is all mixed. There is no a threshold applied
in the picture. We need to do a further process.

Figure 5. Numbered fingerprint segment

In Figure 5 we see there are nine segments. For example, in
segment number 1, there are many red pixels. Those indicate
the value will be used in Hill Cipher. Imagine that we have 400
dots in segment 5. This value has to face the modular
expression to make the number is suitable in Hill Cipher
Process. Modulo also adjusts the determinant to give the correct
value. The following formula explains how it works.
=

(1)

Where :

Figure 3. The original fingerprint

N
TP
TC

:
:
:

Integer Number
Total pixel in a specific segment
Maximum value of modular expression

The total pixel is obtained from how many red colors in the
segment. The computer program calculates the red pixel one by
one and store to the variable. It changes the pixels to characters.
We have decided to change the white color to . while the red
to X. If we see the following figure, it is like a magnification.
........................................
........................................
........................................
........................................
........................................
........................................
........................................
........................................
........................................
...................................X.XX.
......................................XX
....................................X...
........................................
..............................X......X.X
...............................XX..X.XXX
...................................XXX..
.....................................X..
............................XX.X........
................................XX.....X
..............................X.XX.....X
.................................XXX.XXX
........................X.........XX.XXX
......................XXXXX.......XX....
............................XXXXXXX.X...
................................XXX.XXXX
.......................X.............XXX
....................X..X..XX....X.X.....
.........................X......X....XXX
............................X..XX....X..
...................XXXXX...XX..XX..XXXXX
....................XXXX...X.XXXXXXXXXXX
................XX...X........XX.XXX.XXX
.....................X.XX.......XXX....X
......................XXXXXXXXXXX.XXXX.X
.............X.XX.XX.XX.XX.X.XXXXXXXXXXX
.............XXXXXXXXXXX....XXXX.XXXXXXX
............X............XXXXXXXXXXX..X.
..........................XXXXXXXXXXXX.X
...............XX..XX.XXX..XXXX.XXXXXXXX
.......X...XXXX....XX.XX..XXXXX..XXX..XX
.................XXXXX.......XXXXXXX....
....................XXXXXXXXXXXXXXXXXXXX
...................XXX.XXXXXXXXXXXXXXX..
.............XXX..XXX.XXXXX...XXX.XXX..X
.......X.XX.XXXXX....XXXXXXXXX.XX.XX.XXX
.................XXXXXXXXX.XXX..XXXXX.XX
.........XX...X......XXXXX.XX...XX..X.XX
.........XX.XXXXXXXX..XXX....XXXXXXX.XXX
.............XX..XXXXXXXXX.XXXXXXXXXXXX.
.............X....X.X...XXXXX.X.XXXXXX..
..........XXXX.X...X..XXXXX...XXXXXXX..X
........X......XXXXXXXXXXXX...XXXXXXX.XX
........X........X.XXXXXXXXXXXX.X.XXXXXX
.........X..XXXXXX...XX..XXXXXXXXX...XXX
....X.......XXXXXX.XXX.XXXXXXX.XXXXXXXXX
..................XXXXXXXX.XX.X.XXXXXX.X
.........XXX.XXXXXXX..XXX.XXXXXX.XXX...X
.............XXXXXXX.XX.XX.XXXXXXXXX.XXX
........XX...X....X.XXX.XXX.XX...XXXXXXX
........XX.XXX.XXXXXXX..XXXXXX..X.X.XXXX

Figure 6. The data extracted from the first segment

Figure 6 illustrates the data is retrieved from the first

segment. Since it is at the edge of the finger, the shape is round
as well. We count how many X in the figure. The value is
inserted into the cell respectively.
IV. EVALUATION
The best way to prove the method is through a test. Assume
we have to find a set of Hill Cipher key. There are three
columns and rows in every key. It means we have to complete
the step for nine times. Table 1 below shows that the matrix
model before insertion. X11 to X33 are the cells where the integer
numbers are located.
Table 1. Hill Cipher key.

X11
X21
X31

Key
X12 X13
X22 X23
X32 X33

The experiment runs with three kinds of threshold. The first

threshold is 50, the second is 125 and the last is 200. The three
pictures give the different shape.

Figure 6. (a) Threshold = 50, (b) Threshold = 125,

(c) Threshold = 200
The pictures in Figure 6 have the different display. If we
calculate the red pixel in them, it will show the different result
as well. The set of the first threshold (50) can be seen in Table
2. From the fingerprint, it produces 204, 242, 227, 166, 248,
223, 112, 220 and 84.
Table 2. Key of Threshold 50

204
166
112

Key
242 227
248 223
220 84

Determinant
-1103880
248

The set of the second threshold (125) is in Table 3. The

fingerprint calculation produces 124, 84, 91, 214, 209, 105, 241,
29 and 39.
Table 3. Key of Threshold 125

124
214
241

Key
84
91
209 105
29
39

Determinant
-1961133
83
The set of the first threshold (200) generates 77, 242, 84, 163,
239, 241, 34, 235 and 220. It is in Table 4.
Table 4. Key of Threshold 200

77
163
34

Key
242 84
239 241
235 220

Determinant
-4472371
205

We have three determinants. Remember, not all determinant

work. Why? It is only useful if the computer generates the odd
number. Those determinants show two sections. The real
determinants are -1103880, -1961133 and -4472371 as showed
in Tabel 2 to Table 4. We normalize the numbers by modulo
them with the ASCII length; that is 256. The numbers will turn
into the positive number. The determinant can be the odd and
even number. Since it produces the even number, the value can
return the plaintext. It will fail if we do the decryption process.
From those tables, there are only two sets of keys can be used
as the encryption keys. There are in Table 3 and Table 4. Those
numbers are showing the odd number, 83 and 205. If we force
to decrypt the even number, the computer will face the forever
repetition.
We do not include the encryption and decryption process
because our focus is only to generate the key using the
fingerprint pattern. We assume the readers have been familiar
with the Hill Cipher algorithm. The fingerprint image can
simplify the distress of the key store. We do not have to
remember such numbers and just keep the picture and the
threshold value when doing the encryption and decryption
process.

V. CONCLUSION
There are many ways to construct the key in Hill Cipher. One
of the techniques is using an image and threshold. The numbers
are adjustable to get the correct determinant. The fingerprint
has a plurality of segments. It has a pattern which consists of a
collection of dark and white pixels. We can utilize this situation
to produce the Hill Cipher key. The threshold can be set as
desired to generate the specific numbers. Afterward, the
numbers can be inserted into the Hill Cipher cells. By providing
the correct threshold, the decryption process can be carried out.
REFERENCES
[1] A. P. U. Siahaan, Three-Pass Protocol Concept in Hill
Cipher Encryption Technique, SNATI, Yogyakarta, 2016.
[2] A. P. U. Siahaan, RC4 Technique in Visual Cryptography
RGB Image Encryption, SNATIA, Semarang, 2016.
[3] L. H. Thai dan N. N. Tam, Fingerprint Recognition Using
Standardized Fingerprint Model, International Journal of
Computer Science, vol. 7, no. 3, pp. 11-17, 2010.
[4] S. S. Mudholkar, P. M. Shende dan M. V. Sarode,
Biometrics Authentication Technique For Intrusion
Detection Systems Using Fingerprint Recognition,
International Journal of Computer Science, Engineering
and Information Technology, vol. 2, no. 1, pp. 57-65,
2012.
[5] B. Forouzan, Cryptography and Network Security,
McGraw-Hill, 2006.
[6] H. Anton dan C. Rorres, Elementary Linear Algebra,
2011: John Wiley & Sons.